Effective programming of a photonic processor with complex interferometric structure

Автор(ы)
I. V. Kondratyev, K. N. Urusova, A. S. Argenchiev, N. S. Klushnikov, S. S. Kuzmin, N. N. Skryabin, A. D. Golikov, V. V. Kovalyuk, G. N. Goltsman, I. V. Dyakonov, S. S. Straupe, S. P. Kulik

Physical Review Applied, 25, 3, 034072 (2026)

Аннотация

Reconfigurable photonics have rapidly become an invaluable tool for information processing. Light-based computing accelerators are promising for boosting neural-network learning and inference [X. Xiao et al., APL Photonics 6, 126107 (2021); X. Meng et al., Nat. Commun. 14, 3000 (2023); J. Cheng et al., Nat. Commun. 15, 6189 (2024)] and optical interconnects are foreseen as a solution to the information transfer bottleneck in high-performance computing [Y. Li et al., in 2021 58th ACM/IEEE Design Automation Conference (DAC) (IEEE, 2021), pp. 931–936; A. Rizzo et al., IEEE J. Sel. Top. Quantum Electron. 29, 1 (2022); A. Netherton et al., Photonics Res. 12, A69 (2024)]. In this study, we demonstrate the successful programming of a transformation implemented using a reconfigurable photonic circuit with a nonconventional architecture. The core of most photonic processors is an MZI-based architecture [M. Reck et al., Phys. Rev. Lett. 73, 58 (1994); W. R. Clements et al., Optica 3, 1460 (2016)] that establishes an analytical connection between controllable parameters and circuit transformation. However, several architectures [R. Tang et al., IEEE Photonics Technol. Lett. 29, 971 (2017); M. Y. Saygin et al., Phys. Rev. Lett. 124, 010501 (2020)] that are substantially more difficult to program have improved robustness to fabrication defects. We use two algorithms that rely on different initial datasets to reconstruct the circuit model of a complex interferometer, and then program the required unitary transformation. The first method is based on the global fitting of the experimental calibration data, while the second method is an ML-based approach introduced in Kuzmin et al. [Opt. Express 29, 38429 (2021)]. Both methods performed accurate circuit programming with an average fidelity greater than 99% and 97%, respectively. Our results provide a strong foundation for the introduction of nonconventional interferometric architectures for photonic information processing.

Файл